<-/rf=s{(^o 



168 Comparison of Mercury and Platinum Thermometers. 



The conversion from the platinum to the air scale is made 

 by the following formula : — 



Y- — \ 



) 100 J 



where t is the temperature on the air scale, pt the platinum 

 temperature as defined above, and 8 a constant. (For the 

 experimental evidence in favour of this formula, see the 

 elaborate comparison of the air and platinum thermometers 

 by Callendar and Griffiths, Phil. Trans. Roy. Soc. vol. clxxxii. 

 (1891), A). This relation being a parabolic one, only three 

 temperatures are necessary for the complete standardization 

 of a platinum thermometer, i. e. for the determination of 8. 



■p 

 The identity of our value of ^ with that obtained by Griffiths 



in his standardization of our thermometer, made it unnecessary 

 for us to determine the resistance of the thermometer in 

 boiling sulphur (the third point usually employed), and we 

 therefore accepted the value of S, 1*491, given by Griffiths, 

 as correct. 



All auxiliary thermometers used for resistance box, measure- 

 ment of air temperatures, &c, were compared and reduced 

 to air scale. 



Each of the mercury thermometers was separately com- 

 pared with the platinum thermometer in two independent 

 series of observations. 



The results of the comparisons are best shown by the 

 accompanying curves, in which abscissae represent tempera- 

 tures on Rowland's air scale and ordinates the corresponding 

 corrections which must be added to Rowland's scale to 



6166. 



6163. 



reduce to the air scale obtained through the platinum 

 thermometer. 



The differences in the above comparisons are far greater 

 than can be accounted for by experimental errors, and must 



